Collecting carbon data from the supply chain: running a programme that works

Most Scope 3 data programmes start the same way and stall the same way: a questionnaire goes to 300 suppliers, 40 reply, 15 replies are usable. The problem is not the questionnaire but the absence of segmentation.

Segment first

Classify suppliers on two axes: emissions contribution and relationship strength. High-contribution, strong-relationship suppliers are the primary target — one-to-one engagement pays off there. High-contribution, weak-relationship suppliers need procurement leverage. For low-contribution suppliers, average data is entirely sufficient.

The segmentation can be derived from a first-year spend-based baseline. The purpose of a baseline is not to find the right number but to tell you where to look.

Staged requests

Asking for a verified product EPD at first contact usually goes unanswered — the other side does not know what is being asked or what it costs. Staging the request lifts response rates substantially.

• Stage 1 — mass per product and basic material composition. Almost every supplier can provide this.
• Stage 2 — production site location, energy source, production route. Technical but obtainable.
• Stage 3 — a calculated carbon footprint per product, with a method statement.
• Stage 4 — an independently verified EPD or ISO 14067 report.

Even Stage 1 and 2 data beats a spend-based calculation by a wide margin: knowing material composition and production route lets you pick a much better-matched background database activity.

Building supplier capability

Most small and mid-sized suppliers have nobody who can produce a carbon calculation. Asking for data is not enough; you have to show them how to produce it. Joint training through a sector association, a shared calculation template or platform access opened to suppliers are the interventions that raise response rates most.

That investment has a second return: a supplier able to produce its own product EPD offers the same data to its other customers and starts treating it as a competitive advantage. A relationship that began as an obligation turns into mutual benefit.

Screen incoming data before accepting it

• Unit consistency — kg CO₂e per kg, per item, or per m²?
• System boundary — cradle-to-gate or cradle-to-grave?
• Period — which year’s production
• Method — which standard, which background database, which LCIA method
• Order-of-magnitude check — compare against the literature range for similar materials

The last item catches the most errors. A reply of 0.2 kg CO₂e per kg for a metal part is almost certainly in the wrong unit or on an incomplete boundary. A simple order-of-magnitude check stops most errors at the door.